Check out the latest model drops and powerful integrations.
Check out the screen recording. In video-to-video mode, It's doing an awesome job of tracking my head / embodying the lion, but then at the end it interprets my hand as a human hand. Is there a way to get around this?